Rethinking the Sutton Hoo Shoulder Clasps and Armour Noël Adams The kingly Anglo-Saxon ship burial in Mound 1 at Sutton Hoo is found on the reverse of the clasps, examining parallels and the the richest and the most intact of all surviving barbarian implications of different object types fastened in this manner. assemblages from the Early Medieval period. The grave goods Finally, as the clasps are taken to represent armour or armour- range in date from the late 5th century through the first related fittings, Part III re-evaluates the evidence for defensive decades of the 7th century and current scholarship tends to body armour in the grave, including the actual mail as well as identify the wealthy burial with Raedwald, king of the East the hypothetical ‘imperial’ cuirass. Angles who, according to the later historian Bede, ruled from c. 616 to c. 625 as overlord of the English kingdoms.1 The grave is PART I noted particularly for its assemblage of gold and garnet cloisonné ornaments, of which two pieces, the shoulder clasps, Burial deposition and description are the primary focus of this paper. The Sutton Hoo Mound 1 clasps were found in the grave at The majority of the gold ornaments at Sutton Hoo – the roughly the same level as the sword hilt, a half-metre or so west buckle, sword and scabbard fittings, belt mounts and purse lid of the belt fittings (Pl. 2:1).7 Although the excavators had a – find parallels within the Germanic traditions which evolved clear sense that the helmet, sword and regalia (clasps, gold in Europe, in some cases from 5th-century archetypes. The buckle, belt mounts and purse lid) were distributed in a fashion shoulder clasps, however, have no clear typological precedents, that suggested a body, no body was found.8 When first remaining at once the most refined and the most enigmatic of discovered, the excavator Charles Phillips identified the Sutton the gold regalia preserved in Mound 1 (Pl. 1).2 Hoo clasps as ‘armlets’;9 later Gamber and Bruce-Mitford In his magnum opus on Sutton Hoo Bruce-Mitford argued that their curve was most suitable for wear on the illustrated the cuirass on a 1st century bc imperial portrait of shoulder,10 hence the sobriquet by which they are still known. Augustus (the Prima Porta statue) as a possible model for the The clasps were found pinned together and close to one clasps (Pl. 18 and Part III below).3 Despite his careful phrasing another with one clasp on its edge and the other flat and face as to the relationship between the clasps and statue,4 this up (Pl. 2:2).11 On the 1939 field photographs, both pin heads allusion to imperial armour continues to influence scholarly are facing the same direction. Bruce-Mitford assumed, I believe ideas about both the nature of the clasps and the concept of correctly, that when the clasps were fastened in place on the imitatio imperii which, it has been argued, is expressed by the shoulders the pins could only have been inserted from the grave goods in Mound 1 at Sutton Hoo.5 Immediate objections, outer edges. If buried in this position on the body the pinheads however, may be raised to this parallel. The method by which should have been facing in opposite directions. This the clasps were attached (by small loops on the reverse), for discrepancy, and their close proximity (3.8cm apart), suggests example, would not be compatible with Greek and Roman their stance in the grave does not correspond to their original metal cuirasses and the possibility that the clasps simply position on the shoulders (presumably at least 10–20cm apart). copied a representation of an imperial Roman cuirass type6 is dubious on both chronological and technological grounds. Bruce-Mitford’s analysis of the function of the clasps was brief and, as we shall see, dilatory. The reasons for this are good ones, and many fundamental questions regarding these unusual objects may never be answered. Chief among these is the material to which they were attached and, although some possibilities may be eliminated, the original material will probably never be known. Despite this, the fact that the primary purpose of the clasps does seem comparable to the shoulder fastenings shown on classical cuirasses must be accounted for. A modern reassessment of the clasps, therefore, particularly with regard to their relationship to armour, seems in order. The following discussion is divided into three sections. Following a résumé of the archaeological context of the clasps, Part I reviews the specific techniques used to create the clasps (cloisonné, millefiori and filigree), followed by a brief discussion of the iconography and possible prototypes for the form of the clasps. Part II focuses upon the attachment loops Plate 1 Shoulder clasps, Mound 1, Sutton Hoo (PE 1939,1010.4-5), W: 5.4cm ‘Intelligible Beauty’ | 83 Adams Plate 2:1 View of shoulder clasps in the ground Plate 2:2 Detail of shoulder clasps as excavated Plate 3 Details of shoulder clasps 84 | ‘Intelligible Beauty’ Rethinking the Sutton Hoo Shoulder Clasps and Armour Other elements of the regalia, however, were also in these specifically represent a Late Antique or Byzantine ‘unnatural’ positions due to shifting or disturbance within the tradition is worth reviewing, although the issue must remain grave.12 unresolved. In Volume I of the British Museum publication of Mound 1, Complete stepped rhomboid plates and patterns first Bruce-Mitford hypothesised that ‘the shoulder-clasps, detached appear on elite sword fittings deposited in the middle decades from the garment to which they had once been fixed, had been of the 5th century. The distribution of these is, with one placed together in the middle of an empty space where the exception, wholly western,19 but Arrhenius’ hypothesis that the body should have been’, admitting that the ‘placing of the seax scabbard mouthpiece with linked stepped rhomboids shoulder-clasps... remains unexplained’.13 Although his text found in the burial of King Childeric at Tournai (d. 481) (Pl. vacillates as to what the clasps may have actually fastened or 4:1) was made in Constantinople has been influential on recent have been fastened to,14 he personally felt they must have been scholarship.20 I have argued that such plates may have been fitted to a cuir bouilli cuirass.15 In light of the small attachment developed as an expedient means of decorating multiple loops on the reverse, researchers have allowed for the weaponry fittings, not impossibly in association with the possibility of textile in the equation, and most recently have fabricae of the Roman Empire,21 but, as the manufacture of doubted altogether the possibility of a hardened leather arms and their decorative fitments could have taken place in cuirass.16 different places, at present there is no way of determining where such cloisonné workshops may have been located, or The cloisonné, form and motifs indeed whether the goldsmiths were fixed or peripatetic. Each clasp consists of two symmetrical curved halves pinned Weapons with this particular cloisonné cellwork found their together by meticulously engineered hinges, still in way into Hunnic hands as well (Pl. 4:2)22 – whether this functioning order. The sophisticated design of each section – suggests production by jewellers working for clients on the panels of stepped rhomboids, Germanic Style II interlace and Danube during the Hunnic occupation of Pannonia or simply crossed boars – is matched by flawlessly executed garnet demonstrates the trajectory of trade and control from the cloisonné, millefiori glass and zoomorphic filigree (Pl. 3). Eastern Roman Empire likewise remains unknown. Analysis of each of these elements in a meaningful way is Stepped rhomboids do not at present feature on the beyond the scope of this paper but, as they were not cloisonné-ornamented swords with gold hilts found in investigated in depth by Bruce-Mitford, the following Merovingian graves in the period from c. 480–520/25, which, it comments provide some background and dating evidence. has been argued, are local imitations of ‘Byzantine’ The ‘carpet’ of regular stepped cellwork on the main panels prototypes.23 The immediate predecessors to the shoulder clasp of the clasps may appear a rather straightforward geometric decoration are found in the next phase of production, on pattern, yet this is one of the most notable features of the cloisonné sword pommels.24 The earliest of these, a pommel clasps. The plate shape known as the ‘stepped rhomboid’ has with single stepped rhomboids on both faces, was found in the been the focus of much discussion,17 but overall patterns rich grave of a Frankish leader at Krefeld-Gellep on the lower assembled with these plates are relatively rare in the corpus of Rhine (Grave 1782), deposited in the second quarter of the 6th garnet cloisonné. This reflects to some degree the lapidary century.25 Two other examples, from Stora Sandviken (Stürko, capabilities of different jewellers but also the availability of Blekinge) (Pl. 4:3) and Väsby (Hammarby, Uppland) in adequate numbers of well-matched stones,18 which tended to Sweden, incorporate patterns of linked stepped rhomboids; be restricted to goldsmiths working for high-status clients. In these were made and deposited, on current evidence, in the the context of the conference, the question of to what extent period from 560/70 to 600/610.26 Arrhenius argued that the Plate 4:1 Seax scabbard mouthpiece, tomb of Childeric, Tournai, Belgium; Plate 4:2 Sword mouthpiece and scabbard, Voshod (Engels-Pokrovsk), Russia; L: 7.3cm L: approx. 9.5cm Plate 4:3 Pommel, Stora Sandviken, Sweden; L: 7.3cm Plate 4:4 Buckle, Ostrogothic period; L: (buckle-plate) 5.9cm ‘Intelligible Beauty’ | 85 Adams Plate 5:1 Pair of rectangular belt mounts, Mound 1, Sutton Hoo (PE Plate 5:2 Belt mount, Grave 7, Rödingen, Germany; L: 5.2cm 1939,1010.8-9); L: 5.2cm Plate 5:3 Belt mounts, Grave 127, Schretzheim, Germany; L: (largest) 5cm Plate 6 Buckle counterplate, Kerč, Crimea; H: 4.2cm Plate 5:4 Belt mounts, strap distributor and buckles, Mound 1, Sutton Hoo 86 | ‘Intelligible Beauty’ Rethinking the Sutton Hoo Shoulder Clasps and Armour pommels preserved in Sweden were made in Frankish contemporaneous (see note 31). On the basis of their workshops,27 but pommels of this class have been found in relationship to the Vendel sword pommels and to one pair of Gotland, Finland, England, Bavaria and Italy28 so their place of the belt mounts, it may be suggested that the Sutton Hoo clasps manufacture is uncertain. It is possible that the cloisonné were produced in the period from 560/70–600/610, perhaps, pommels preserved in Scandinavia reflect workshop traditions given the size of the garnets, on the earlier end of this time originating in Italy, where Ostrogothic burial customs frame. prohibited the inhumation of weaponry but where high-status The integration of glass into the ‘carpet’ of rhomboids female jewellery made in the first half of the 6th century recalls Ostrogothic-period traditions, taken to new heights of displays large examples of these plates (Pl. 4:4).29 Stepped luxury with millefiori glass. Bruce-Mitford argued that the rhomboids were also used on the famous gold liturgical paten, chequered red, blue and white blue millefiori implied the buried at Gourdon (Saône-et-Loire) sometime after 525,30 so presence of Celtic craftsmen,39 but scientific examination of the jewellers producing or using these plates also worked for millefiori inlays in the British Museum Research Laboratory ecclesiastical clients. The current picture, therefore, suggests established that they were compositionally different to that stepped rhomboids were relatively rare and restricted to contemporary Celtic millefiori glass.40 The canes include a rare high-status ornaments throughout the 6th century. The maker translucent red glass coloured by manganese, similar to that of the Sutton Hoo shoulder clasps had access to enough garnet found on glass tesserae used on Roman wall mosaics, and there to prepare both large and small stepped plates (16 at approx. is little doubt that they were made using recycled Roman 6.5 x 6.5mm and 14 at approx. 5 x 5mm) of the same general size glass.41 The source and place of manufacture of the millefiori as those on the pommels preserved in Sweden. canes, however, remains unknown, although continental Stepped rhomboid garnet plates of the smaller size were Europe must be considered a strong possibility.42 also displayed individually on one pair of rectangular mounts Framing the stepped garnet and millefiori glass rhomboids from the sword belt found in Sutton Hoo Mound 1 (Pl. 5:1), on on the inner rectangular panels of the Mound 1 clasps are the dummy buckle (Pl. 5.4) and are found on at least two interlaced Germanic Style II zoomorphs, inlaid in garnet and cloisonné pommels from the Anglo-Saxon hoard recently blue glass against a gold ground. Like the boar figures, these discovered in Staffordshire.31 The two Sutton Hoo belt mounts are defined by garnet plates freely-cut to individual shapes. share another feature with the clasps – cells filled with gold or The interlaced snaky creatures with large eyes find parallels in covered by a sheet of gold (see below), used on these in garnet cloisonné found in Anglo-Saxon England and Frisia.43 combination with elbow-shaped plates to create a rope-twist The components of Anglo-Saxon Style II have never been interlace around the border. Belt mounts of this form, generally subjected to rigorous dating analysis, but it is clear that the made in gilded copper-alloy with inset nielloed silver panels, shoulder clasp zoomorphs can be associated with the first are useful as dating indicators. Borders of simple rope twist phase of Style II, Høilund Nielsen’s Phase D (565/80–610/30).44 interlace, for example, appear on rectangular belt mounts of Speake noted some general stylistic similarity with the nielloed Type Weihmörting found in a rich Frankish chamber grave at creatures on the silver harness mounts from Grave 9 at Rödingen (Nordrhein-Westfalen, Grave 7)32 (Pl. 5:2) and in Niederstotzingen in Bavaria and the placement of the eye at the Grave 127 in the large Alamannic row-grave cemetery at top of the head may also be compared to the zoomorphs on the Schretzheim (Kr. Dillingen a.d. Donau) in Bavaria (Pl. 5:3).33 Crundale Down (Kent) pommel and Allington Hill (Cambs.) Both belt sets included mounts with boar heads at the corners; escutcheons;45 these, however, lack the angled jaw which is a as Koch noted, both the rope interlace and boar decoration are prominent feature of the Sutton Hoo material and which has rare.34 Grave 127 in the Schretzheim cemetery belonged to the Scandinavian parallels. phase of deposition dated from 565–590/600; Rödingen Grave The technique in which the interlaced zoomorphs are 7 has recently been assigned to phases 6 and 7 of the cemetery executed, relying upon large and small areas of gold sheet to (c. 570–610).35 realise the pattern, was described by Bruce-Mitford as ‘lidded’ The other Mound 1 belt mounts, sword belt strap or ‘beaded’ cells; the device appears on other Anglo-Saxon, distributor, buckle and dummy buckle are decorated with Frisian, Frankish and Vendel Swedish garnet cloisonné.46 interlocking cloisonné cellwork of a somewhat different nature Cloisonné interlace executed in this manner, in the form of (Pl. 5:4). Patterns similar to these, using clusters of elbow and cable twist, also appears on a buckle counterplate from Kerč in mushroom-shaped cells to create focal quatrefoils, however, the Crimea, originally set with garnet plates (Pl. 6).47 This can also appear on objects deposited in the last third of the 6th be dated to the last decades of the 6th or first decades of the 7th century, notably on iron belt buckles decorated in Tauschierung century, a reminder that what Bruce-Mitford took to be a rare (metal inlaying) found in high-status Alamannic and Frankish and specialised western technique was also practiced in the warrior graves.36 It has been suggested that the introduction Byzantine-influenced East. and spread of the technique of inlaying silver and gold into iron Finally, the zoomorphic filigree found in the interstices in Europe was due to the transmission of Byzantine objects between the boars’ legs, like the Style II zoomorphs, finds through Lombardic Italy,37 but it is unknown whether the parallels within insular and Scandinavian traditions.48 The decorative patterns were also imported or whether existing designs of the filigree differs on the two clasps; like the subtle cloisonné objects were being copied. Stepped rhomboids do not differences in the length of the sections of the hinges, this was at present appear on Type Weihmörting belt mounts and are clearly deliberate and in keeping with ancient and medieval rare in inlaid ironwork.38 Recent garnet cloisonné finds from jewellery production which tended to regard perfectly matched Staffordshire, England, however, confirm that patterns pairs as inauspicious. incorporating small mushroom and small rhomboid cells were ‘Intelligible Beauty’ | 87 Adams Plate 7 Vessel mount, Grave 119, Kölked-Feketekapu B, Hungary; H: 8.7cm Plate 8:1 Muthmansdorf type belt mounts, ?Dunapatele, Hungary; L: (of mount with eagle Plate 8:2 Muthmansdorf type belt mount, ?Rome, heads) 5.1cm Castellani collection; L: approx. 11.5cm The primary motif of the shoulder clasps is, of course, the head motifs, often in association with double eagle heads and magnificent figures of crossed boars on the arched panels. sometimes the heads of hounds as well (Pl. 7),54 were a key part These incorporate large chequerboard millefiori inlays at their of the repertoire of high-status jewellery and ornaments used shoulders and include some of the largest garnet plates found by the Germanic elites in Europe, England and Scandinavia in on the Sutton Hoo jewellery. The boars undoubtedly had the later 6th and first half of the 7th centuries. Individually particular significance to their owner, which we can only these animals were associated with Germanic deities, but their attempt to define. In Britain the boar often had an expressly combination derived from classical hunting imagery and their military significance which can be traced over centuries, from conjoined form evolved from ancient steppe conventions.55 its use as the symbol of the Legio XX Valeria Victrix (which In comparison with many of these, the complete boars on invaded Britannia in ad 43 with Claudius)49 to the actual boar the Mound 1 clasps are, given the constraints of cold cloisonné tusk ornaments found at the Late Roman fort at Richborough50 and the space into which they are compressed, remarkably to the later 7th-century boar-crested helmets preserved at realistic. Yet their crossed position, with their hind legs and Benty Grange and at Woolaston.51 snouts touching the ground and front legs dangling in the air, In the late 6th and early 7th centuries the cloisonné sword contrasts with both Roman and medieval images of boars, pommels found in Sweden incorporated cryptic which typically show the beasts running, leaping or seated. representations of boars and hounds (Hög Edsten, Bohuslän) Possible interpretations of this position are that the boars are and boar heads (Vallstenarum, Gotland) and a mid-6th- fighting, dead 56 or that boar skins are represented.57 Whether century continental silver and niello pommel included in the seen as battling animals or trophies of the hunt, these boars Staffordshire Anglo-Saxon hoard is decorated on one side with have been transformed into a specific visual signal. Such an addorsed zoomorphs with boar heads at the corners.52 The emblem might have been purely personal, tribal or may have complete crossed boars on the shoulder clasps remain had a more formal role as an ‘official’ insignia. unparalleled, although the paired treatment is in keeping with These lines of reasoning acquire further relevance when we imagery of the period. Small copper-alloy mounts in the form examine the unusual overall shape of the clasps – a rectangular of conjoined boars, for example, were used on seax belts found panel surmounted by an arch. This form can be paralleled on a in Alamannic and Frankish graves from the last third of the 6th class of Late Roman belt mounts known as the Muthmansdorf century through the first two decades of the 7th century.53 In Type, after a find in Austria.58 Mounts of this type were addition a wide range of fittings decorated with paired boar engraved with panels of sophisticated figural imagery in 88 | ‘Intelligible Beauty’ Rethinking the Sutton Hoo Shoulder Clasps and Armour combination with chip-carved ornament. Mounts said to have leather or textile.68 The clasps weigh 183.80 and 201.57g, been found in Dunapentele, Hungary, depict boar hunts however, and would be difficult to secure with anything other between crosshatched panels (Pl. 8:1) while an exceptionally than stout cord. large example, said to have been found outside Rome, In this instance much later 17th- to 19th-century ornaments incorporates borders of cable-twist interlace and addorsed with loops on the reverse, widespread from Central Asia and griffin heads (Pl. 8:2).59 Others of this type are decorated with the Caucasus to Hungary and Scandinavia may offer some circus motifs, pagan and Christian figural themes.60 The figural guidance. Although rows of looped ornaments could be style and type of portrait medallions on these mounts date the secured by a single thin leather thong or textile cord,69 other Muthmansdorf type to the last third of the 4th and first decades looped mounts were traditionally secured by individual short of the 5th century.61 strips or splints of leather.70 These were inserted when the Unlike other chip-carved belt sets found in the graves of leather was fresh and malleable so that as they dried and Germanic officers serving in the Late Roman army, no hardened the ends expanded and held the object securely in Muthmansdorf mounts have been excavated in situ in a grave.62 place. Examples are distributed along the Rhine and Danube As we shall see, multiple attachment loops on the back of frontiers, with outliers in Italy, Anglo-Saxon England and objects can be documented throughout the 1st millennium on North Africa and we know nothing of the places of production high-status gold ornaments. Although single long loops were of these large and sophisticated figural mounts.63 Some seem to commonly used to fasten gold mounts sleeved with shell collars be personalised with portraits of a helmeted solider and a onto continental and Anglo-Saxon scabbards,71 multiple woman (?his wife)64 and it is possible that they had reverse loops for attachment are much rarer. The great connotations of civil as well as military rank in the Late Roman majority of metal fittings made in England in the late 6th and bureaucracy. These buckles and their matching counterplates 7th centuries followed the established Roman and Germanic provide one possible model for the form as well as the motifs of norm of mounting with dome-headed rivets. By the early 6th the Sutton Hoo clasps, with the boar hunting motifs condensed century in the Byzantine East such riveted systems were and stylised, the geometric panels of cross-hatching reinvented replaced by integral cast lugs, pierced at the end; these became as stepped rhomboids and the guilloche borders transformed the primary method of attachment on belt buckles and other into zoomorphic interlace. ornaments. Lugged mounts begin to appear on belt buckles in I have argued elsewhere that the revival of Late Roman continental Europe and Anglo-Saxon England in the later 6th forms formed a key part of the early Anglo-Saxon aesthetic century.72 The rarer looped attachments appear briefly in the response to their Germanic and Roman past.65 As for the West on Late and post-Hunnic period ornaments in the 5th relevance of animals of the hunt (if this is what the boar motifs century and then re-occur in increasing numbers from the indeed signify) to a high-status warrior in the later 6th century, second half of the 6th century. we need only to turn to the Early Byzantine military manual, Small attachment loops were originally a feature of steppe the Strategikon, where hunting from horseback is described as nomadic metalworking, but were in use in western a complementary discipline to warfare, keeping men alert, goldsmithing at least as early as the Hellenistic period.73 In the exercising their animals and teaching them tactics.66 case of precious metal fittings, round or strip wire was folded Although we might propose that the form and decoration of into a loop whose ends were then soldered to the backs of the Sutton Hoo clasps may be traced back to Late Roman decorative mounts or functional buckles and clasps. Once in official belt sets, the archaeological evidence does not allow us place objects held by multiple loops were very secure, but at the a clear understanding of such an evolution, or functional same time this system, unlike the riveted one, permitted transformation. Some further insight into the development of these mounts is offered by an examination of the loops on the backs of the clasps, a neglected feature which finds parallels in high-status Byzantine-period objects in this period. PART II Looped versus pinned attachment systems The Sutton Hoo clasps were affixed by means of narrow loops of 1.2mm gauge wire on the reverse; each pair has 20 of these, arranged in parallel rows. Each loop measures approximately 3mm in length with an internal opening of approximately 2.0– 2.5mm (Pl. 9).67 The workable opening of the loops is so narrow that modern mounting pins measuring 0.5mm in diameter, when sleeved with plastic, fit snugly within them. The slight ‘halo’ visible around the loops on the reverse of the central rows of one of the Sutton Hoo clasps may be evidence that the loops had individual fastenings and Bruce-Mitford apparently considered the possibility that each loop could have been secured by a pin, but, as nothing of this sort survived in metal, concluded that they must have been sewn into place against Plate 9 Shoulder clasps, reverse and side views; W. 5.4cm ‘Intelligible Beauty’ | 89 Adams Plate 10:1 Belt mount, Kunágota, Hungary; H: 4.2cm Plate 10:3 Strap mount, Grave 90, Castel Trosino, Italy; H: 4.4cm Plate 10:2 Buckle, Sirmium (Sremska Mitrovica), Republic of Serbia; L: 8.1cm greater flexibility in elite dress, allowing ornaments to be distributed across Byzantium, Sasanian Persia and the added to an existing garment or belt, or removed, for cleaning, northern steppe zone where mounts with multiple loops were repair or transfer to another outfit. The following discussion also being made.86 They are not, however, imports, but native reviews the three primary types of functional fittings with Lombardic variations, in some cases manufactured in the reverse attachment loops: buckles, hook and eye clasps and Preßblech technique using dies which copy eastern hinged clasps. prototypes.87 Their Alamannic counterparts found north of the Alps were for the most part made in baser metals or were Buckles constructed with riveted pins.88 Further north, in core Although a cloisonné buckle with loops on the reverse was Merovingian territories, loops can be documented only rarely, found in a Late Sarmatian period grave of the second half of on high-status mounts such as those excavated in Grave 11 at the 3rd or early 4th century,74 the majority of surviving buckles Saint-Denis.89 and belt mounts with loops appear in contexts dated to the In Anglo-Saxon England another pair of clasps with second half of the 6th and 7th centuries. In addition to high- attachment loops survived in a large, male, barrow grave at status gold buckles of undoubted Byzantine type or origin,75 Taplow in Buckinghamshire (Pl. 11); the grave goods in this variants in silver76 and copper-alloy77 have also survived. Early deposit range in date from the second half of the 6th to the coin-dated examples in the West include the belt set found in early 7th century.90 The character of the burial is not dissimilar an Early Avaric rider’s grave at Kunágota, Hungary with a to that of Mound 1 at Sutton Hoo, although its monetary solidus of Justinian I (545–65)(Pl. 10:1 and Daim this volume, wealth, in terms of the gold deposited, is considerably less. The Pl. 5)78 and a gold buckle in a female grave at Perugia buried Taplow clasps differ from the Sutton Hoo clasps in significant with 140 coins of Justin II with a terminus post quem of 578.79 ways,91 but a brief examination of these and their antecedents In the middle decades of the 7th century elaborate gold establishes another means by which we can begin to unravel buckles and belt mounts with attachment loops are found in possible prototypes for the Sutton Hoo clasps. Middle Avaric and Bulgar contexts of the highest status, for example the khagan’s belts from Kunbábony,80 Sirmium (Pl. The Taplow clasps 10:2)81 and Mala Pereščepino,82 as well as the high-status male The Taplow clasps were found in a textile sandwich and grave at Bócsa.83 Such buckles seem to represent a fusion of orientated, according to the most reliable plan, one above the steppe nomadic traditions with Byzantine-period goldsmithing other alongside the body with the two halves angled together techniques, but not necessarily actual Byzantine but probably not articulated.92 They were attached to patterned workmanship.84 leather and associated with a textile of red tabby wool.93 The Between these benchmarks we find examples of gold and contents of the barrow were greatly disturbed but a few silver sheet buckles and strap mounts with loops (in some cases surviving bones and the position of the grave goods suggested for horse harness) primarily in Lombardic graves in Italy (Pl. the body was orientated with the head to the east94 with the 10:3).85 Some of these reflect types which are archaeologically clasps located on the left-hand side. 90 | ‘Intelligible Beauty’ Rethinking the Sutton Hoo Shoulder Clasps and Armour Plate 11 Front and reverse of pair of hook and eye clasps, Taplow (Bucks) (PE 1883,12-14.2); W: 2.9cm The clasps are made of copper-alloy covered with gold of Georgia) provides evidence that similar garment fastenings sheet. Each clasp has three wire loops on the reverse, were worn on the north-western border of the Parthian Empire positioned at the shoulders and toe (Pl. 11). These were (Pl. 12:2).97 Although this was a female grave, Musche has inserted from the back through the gold sheet and copper-alloy suggested the clasp represents a type of male cloak fitting.98 A plate with the ends hammered down and disguised by the disc-shaped hook and eye clasp in the State Museum of Georgia inlays in the circular cells at the corners.95 Although made of was decorated in the new fashion of garnet cloisonné, probably less robust wire (1mm thick) than the loops on the Sutton Hoo in the later 4th century (Pl. 12:3),99 while an unprovenanced clasps, their opening is actually slighter larger (2.5–3mm). The pair of quatrefoil-shaped clasps in silver has been considered clasps were made with one half terminating in a fixed loop and Sasanian work of the 5th century,100 suggesting continuity of the other fashioned with a hook. Each element resembles a these particular garment fastenings in Persia. triangular buckle plate, yet they must have served a different I know of no evidence of hook and eye clasps in the Late function to that of a buckle. This is indirectly confirmed by the Antique Roman world, unless we entertain Werner’s suggestion orientation of the loops on the reverse of the clasps whose that the small pair of semi-circular mounts with loops on the openings are all orientated in the same plane, in contrast to buckles whose loops (or lugs) were set with the toe loop perpendicular to the others, presumably for greater stability. Functionally therefore the Taplow clasps can be seen as unique variants of hook and eye clasps, an object type in use throughout the 1st millennium which was also typically secured by means of reverse loops. Hook and eye clasps In contrast to the elite male buckles and mounts reviewed above, hook and eye clasps have survived primarily in female grave contexts. Early examples of these excavated in situ in 1st-century royal nomadic burials at Tillya Tepe, Afghanistan, show that they were worn either at the chest or neck to close both long sleeved robes and shorter capes or cloaks worn over the shoulders (Pl. 12:1).96 A lobed hook and eye clasp found in Plate 12:1 Gold disc-shaped clasps, Tillya Tepe, Afghanistan; Diam: 3.3cm a disturbed 2nd-century burial in Tomb 7 at Mcheta (Republic Plate 12:2 Gold, garnet and turquoise clasps, Grave 7, Mcheta, State Museum of Georgia, Tbilisi, Republic of Georgia; W: 5cm ‘Intelligible Beauty’ | 91 Adams Plate 12:3 Gold and garnet cloisonné disc-shaped clasp, State Museum of Georgia, Tbilisi, Republic of Georgia; Diam: 2.1cm reverse from the 5th-century tomb of King Childeric at Tournai techniques and decoration.105 An unpublished clasp in the (Pl. 13, bottom) fastened a sleeveless fur jacket such as that British Museum (PE 1916,2-11.2) has empty cloisonné cellwork which Sidonius Apollinaris described as worn by the comparable to that on the strap end and belt mount from Burgundian Prince Sigismer.101 These lack hooks and eyes but Kunbábony and the strap end from Bócsa,106 i.e. from the may well offer some evidence about garment fashions in the middle decades of the 7th century. The British Museum clasp Late and post-Hunnic periods. Current scholarship views many has no known provenance, but was considered Lombardic of the garnet cloisonné fittings and weapons buried with the when acquired. When excavated in situ, the majority of Frankish king as reflecting Eastern Roman types if not 7th-century Avaric clasps have been found at the jaw, neck and necessarily workmanship.102 Other ornaments in the grave, shoulder, indicating that, they, too, were used to close a such as the bee or cicada mounts and the ox mount, were also garment at the neck or upper chest.107 made with single or multiple loops on the back.103 Clearly loop To return to Anglo-Saxon England, the combined evidence fasteners already formed part of the repertoire of high-status suggests that identification of the Taplow clasps as belt Late Antique goldsmiths in the 5th century, even though there fasteners worn one above the other108 must be reconsidered. is little evidence of these outside this barbarian deposition in The clasps could not be adjusted (a key requirement for a waist the West. belt) so their putative belt would have to have been worn very Hook and eye clasps secured by loops re-emerge in the tightly to maintain horizontal tension on the loops. It is archaeological record in the 7th and 8th centuries in Avaric probable therefore that the triangular buckle in the Taplow and Lombardic material cultures in Europe. A considerable grave served as the primary waist buckle rather than as a number of these have been preserved104 and it has been argued baldric buckle, as its disturbed position in the grave has that the earliest group of Avaric disc-shaped mounts made in suggested. Furthermore, the red-dyed wool tabby fragments the first half of the 7th century imitate both Byzantine associated with the clasps are most convincingly interpreted as a garment rather than a belt. It may be significant here that the hook on one of the Taplow clasps is torqued and damaged. This could be explained if the clasps were subjected to different tension at different levels. This might occur on a central closing of a cloak if it was very tight, but is perhaps more likely on a coat closing with a flap. Long-sleeved overgarments with flaps crossing from right to left, a type which evolved from steppe nomadic riding Plate 13 Colour version of Chifflet 1655 drawing of mounts from tomb of Plate 14 Gold and garnet cloisonné hinged mount or buckle plate, Ureki, State Childeric at Tournai, Belgium Museum of Georgia, Tbilisi, Republic of Georgia; L: 4.8cm 92 | ‘Intelligible Beauty’